INFLUENCE OF HOT WOEKING ON STEEL. 267 may reasonably be assumed that it will give a regular basis of comparison, so that the differences between the results on this standard and on the various thicknesses will be the measure of the effect of rolling. It is shown that for a.n increase of one-eighth of an inch in thickness there is a diminution in strength of 700 pounds per square inch. It is, perhaps, as close an agreement as could be expected when we find that in Table XIY-F the difference on the large sizes between the three-eighth-inch and three-quarter-inch angles, was 1830 pounds per square inch, or 610 pounds to every one-eighth in thickness, while on the smaller sizes it is 2168 pounds from five-sixteenth-inch to five-eighth-inch, or 434 pounds to every eighth, being an average of 522 pounds for both large and small sizes. SEC. XlVg.—Physical properties of Pennsylvania Steel Company steels of various compositions, when rolled into angles of different thicknesses.—The subject is more fully investigated in Table XIV-H, which gives the average results from angle bars of several different kinds of steel. The accidental variations in the metals make it impossible to compare the influence of the thickness upon the ultimate strength, but the column showing the elastic ratio proves that a lower elastic limit follows an increase in thickness. The elongation remains the same for all thicknesses. The reduction of area varies somewhat, but in the groups where a large number of tests make the figures of much value there is a decrease in the heavier bars. The variation in strength of the different thicknesses is due in part to the fact that the thin pieces are finished at a lower temperature. The effect of such working is investigated in Tables XIY-C and XIV-D, where pieces of the same billets were heated differently before rolling and were, therefore, finished under unlike conditions. In the bars finished at the lower temperature the elastic limit was raised very considerably, but the ultimate strength and the ductility did not vary much from the hot-rolled bars. This conclusion has nothing to do with the fact so well known to all manufacturers that if a bar or plate is finished so cool that it looks dark in the sunlight it will give a much higher tensile strength; the bars referred to in the table were all finished somewhat hotter than this, and the small variation in temperature seems to have little effect. These conclusions will be corroborated by Table XIV-I, which records certain tests on acid open-hearth steel.